Re: Pushes, Pops, NOPS

"Barak Zalstein" <Barak.Zalstein@ParthusCeva.com>
8 Dec 2003 00:16:56 -0500

          From comp.compilers
Related articles
Pushes, Pops, NOPS vbjohny@yahoo.com (2003-12-03)
Re: Pushes, Pops, NOPS Barak.Zalstein@ParthusCeva.com (Barak Zalstein) (2003-12-08)
Re: Pushes, Pops, NOPS tmk@netvision.net.il (2003-12-08)
Re: Pushes, Pops, NOPS nick.roberts@acm.org (Nick Roberts) (2003-12-08)
Re: Pushes, Pops, NOPS kenrose@tfb.com (Ken Rose) (2003-12-08)
Re: Pushes, Pops, NOPS vbdis@aol.com (2003-12-08)
Re: Pushes, Pops, NOPS lex@cc.gatech.edu (Lex Spoon) (2003-12-13)
| List of all articles for this month |
From: "Barak Zalstein" <Barak.Zalstein@ParthusCeva.com>
Newsgroups: comp.compilers
Date: 8 Dec 2003 00:16:56 -0500
Organization: Compilers Central
References: 03-12-032
Keywords: architecture
Posted-Date: 08 Dec 2003 00:16:56 EST

> If I added some "NOPS" in between each of the instructions above just
> to make the code look bigger, is there any option while compiling with
> GCC which would eliminate all those unnecessary "NOPS" and then
> convert the remaining code to binary?


GCC treats those kind of __asm__ statements as hard-coded strings, and
doesn't modify them between parsing and final assembly file emission.
A post-compilation pass might be able to eliminate the nops, but I
don't think that current tools utilize that, or find its
implementation worthwhile (it is better to investigate why redundant
nops are generated in the first place).


Barak

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.